Modelling and optimization of a bi-objective flow shop scheduling with diverse maintenance requirements. Issue 9 (3rd May 2018)

• Record Type: Journal Article
• Authors: Seif, Javad; Yu, Andrew Junfang; Rahmanniyay, Fahimeh
• Is Part Of: International journal of production research. Volume 56:Issue 9(2018)
• Page Start: 3204
• Page End: 3225
• Publication Date: 2018-05-03
• Subjects: flow shop scheduling -- age-based maintenance -- multi-objective optimization -- construction machinery management -- maintenance-dependent processing time; Factory management -- Periodicals; 658.57
• DOI: 10.1080/00207543.2017.1403660
• Languages: English
• ISSNs: 0020-7543

Abstract

Abstract : In real-world problems, machines cannot continuously operate and have to stop for maintenance before they fail. Lack of maintenance can also affect the performance of machines in processing jobs. In this paper, a permutation flow shop scheduling problem with multiple age-based maintenance requirements is modelled as a novel mixed-integer linear program in which the objectives are conflicting. In modelling the problem, we assume that infrequent maintenance can prolong job processing times. One of the objectives is to minimise the total maintenance cost by planning as few maintenance activities as possible to only meet the minimum requirements, and the other objective tries to minimise the total tardiness by sequencing the jobs and planning the maintenance activities in such a way that the processing times are not prolonged and unnecessary maintenance times are avoided. Because of this conflict, an interactive fuzzy, bi-objective model is introduced. Application of the model is illustrated through a case study for operations and maintenance scheduling of heavy construction machinery. An effective and efficient solution methodology is developed based on the structure of the problem and tested against commercial solvers and a standard GA. Computational results have verified the efficiency of the proposed solution methodology and show that unlike the proposed method, a generic metaheuristic that does not consider the unique structure of the problem can become ineffective for real-world problem sizes.